Several studies have shown that a low magnesium (Mg) intake in the diet is associated with greater cardiovascular risk and greater risk of diabetes. However, the results are not consistent in all populations. To minimize the biases derived from diet measurement, more objective biomarkers of magnesium status have been proposed. Although there is still no ideal biomarker for Mg, several studies have shown that plasma Mg concentrations could be a relatively acceptable biomarker for cardiovascular risk assessment. However, further studies are required to better characterize this marker in different populations. Our aim was to analyze the association between plasma Mg concentrations (measured through inductively coupled plasma mass spectrometry (ICP-MS)) methods, and cardiovascular risk factors in individuals from a general Mediterranean population (aged 18–80 years). The influence of demographic and lifestyle variables, including adherence to the Mediterranean diet, on plasma Mg concentrations was analyzed. The mean Mg level of the population studied was 0.77 ± 0.08 mmol/L, the prevalence of hypomagnesemia (<0.70 mmol/L) being 18.6%. We did not find any statistically significant differences between plasma Mg concentrations and sex, age, tobacco smoking and total adherence to the Mediterranean diet (p > 0.05). We found a statistically significant association between plasma Mg concentrations and the prevalence of type-2 diabetes (0.77 ± 0.08 mmol/L in non-diabetics versus 0.73 ± 0.13 mmol/L in diabetics; p = 0.009). Despite the low prevalence of type-2 diabetes in this population (11.24% in subjects with hypomagnesemia versus 3.91%, in normomagnesemia; p = 0.005), hypomagnesemia was associated with greater odds of being diabetic in comparison with normomagnesemia (OR = 3.36; p = 0.016, even after adjustment for sex, age, obesity, and medications). On the other hand, no statistically significant association of plasma Mg concentrations with obesity, hypertension, fasting triglycerides, HDL-cholesterol or uric acid was found. However, in contrast to what was initially expected, a statistically significant association was found between plasma Mg concentrations (basically in the highest quartile) and greater total cholesterol (p < 0.05) and LDL-cholesterol concentrations (p < 0.05). In conclusion, our results contribute to increasing the evidence gathered by numerous studies on the inverse association between hypomagnesemia and type-2 diabetes, as well as to the observation, previously reported in some studies, of a direct association with hypercholesterolemia. This paradoxical link should be deeply investigated in further studies.